Power Management IC

[jabroon]

Hello,

Im currently working on a little hobbyist project and Im stuck...

It's a Arduino based door lock. It will be powered over PoE primarily and when it is connected to PoE I would like it to charge a lithium polymer battery. However, when the PoE connection dies I would like it to switch over to the lipo battery.

The door lock needs to run on 5v (that is what the ATmega328 requires, however if I run it through the on-board voltage regulator then it requires 7-12v, but I would prefer it to provide it with 5v straight into the 5v rail rather than running it through a voltage reg) and 12v for the door striker (@ ~ 500ma, but only for 4 seconds at a time).

I have looked at Digikey and multiple manufacturers websites trying to find a solution. Im confused as whether to use an IC that manages switching between the power sources, charging the lipo battery, managing PoE and doing the dc-dc switching (from ~48v from the PoE down to 5v for the Arduino and 12v for the door striker, and however many volts the battery charging circuit will require). Is there an IC that is capable of doing all of that? Is it better to have multiple different IC's to manage each part?

I am trying to do this right, the way than an actual engineer would do it. I'm currently on school holidays (I'm 16 and in year 12) so I have a lot of time and am trying to further my electronics knowledge by building little projects.

Thank You,

Jabroon.

[HalfSpace]

Hi jabroon,

Can you use a small 12v battery (lipo or whatever) run both the Arduino and door striker from it and then trickle charge the battery from the PoE 48v?

Is there a reason why you don’t like using the regulator on the Arduino?

HalfSpace 

[Whales]

(I'm assuming your POE is 5v, otherwise you would have to use the Vreg anyway).

It looks like you are trying to tackle all of the power situations at once.  Break them down into modules and solve them separately.  Eg:

• Connection of PoE -> 5v rail of Arduino
• Circuitry that lets power from the batteries into a VREG only if no PoE voltage is present
• Circuitry that stops power coming out of the vreg from going through the PoE wires (I assume the twisted pair does not become open-circuit when the PoE is turned off.  I may be wrong)
• Circuitry to charge the batteries only if PoE is present

These things can all be hooked up together and you should be able to do it without them interfering with each other.  Make sure you put a largish cap across the micro's power rail, else transient dips in the power supply (caused by switching from one power supply to another) might be too big for the micro to survive without resetting.

[retrolefty]

I would worry about not having a 'fail safe' condition where as the door would be unlocked in case of building power failure, fire, or other natural disaster?

[Marco]

Can the door striker take a little more voltage? What voltage lipo pack do you want to use?

[jabroon]

Thanks for all of the replies!!!

HalfSpace, it seems that most lipo batteries used today are 3.7v per cell. It would probably be best to just have a single cell and boost it up to the required voltage with a dc-dc converter. I read what you wrote as saying that I should connect both the charging circuitry and the load directly to the battery and simultaneously charge and discharge the battery - however I may be wrong, but if that is what you meant then this is why I am hesitant: from what I've seen on the internet, when you both simultaneously charge and discharge a battery (in this case charging the battery from 48v PoE being stepped down to the charging voltage and being discharged by the Arduino and door striker) it does work; however it is seen as a bit of a hack as you are constantly running the battery, degrading its life and performance more rapidly over time. Regarding the Arduino's voltage regulator, the idea of wasting the 7v (12v input - 7v from volt reg to yield 5v for Arduino) sorta annoys me . However, I have been thinking recently () that it would be a good tactic to use a dc-dc converter to do the heavy lifting and bring whatever voltage (probably 12v) down to about 7v and then run it through a voltage regulator and get a much cleaner 5vdc.

Whales, I plan on using PoE in accordance with 802.3af, so it would be running at a nominal 48v. Thats an interesting approach. I was initially going to do that and design a little logic circuit or something and then use some MOSFETS to switch all of the power. I don't really think that it would be too difficult. The thing is that I began to read up on these PMIC's and got swayed a bit, they seem perfect for my situation but I just get lost in their data-sheets. I will make a circuit that should do the job but I would still love to see how to use one of these little power management ic's. Thats a good idea to place a decent capacitors across the power rail (btw, would this be an example of a decoupling capacitor or is there another name for it?).

Retrolefty, your correct in identifying safety as a big thing. There are multiple things that make this system rather safe:
1. the doors only stop people from entering a door, not exiting. So in the event of a power outage or natural disaster or even a complete system failure people can still exit the building.
2. the purpose of the lipo battery is to provide power to the system and give some up-time (potentially quite a bit) so people can get in and collect what they have to in the event of a power loss situation
3. in a situation where security wasn't as important as convenience (for example a bed room door lock) a door striker that opens upon power loss could be used

Marco, the door striker is pretty robust. I guess it could take a lot more voltage, but I don't understand what advantage is gained by overvolting it. A 3.7v single-cell lipo pac is currently my number 1 choice, however, that is subject to change if an advantage can be gained from another solution.

Again, Thank You all for the replies

[retrolefty]

Retrolefty, your correct in identifying safety as a big thing. There are multiple things that make this system rather safe:
1. the doors only stop people from entering a door, not exiting. So in the event of a power outage or natural disaster or even a complete system failure people can still exit the building.

 Like fireman trying to enter to fight a fire? 

But then again fireman do have means to enter a locked door if they deem necessary. But do check with your local code requirements.

[jabroon]

Like fireman trying to enter to fight a fire? 

But then again fireman do have means to enter a locked door if they deem necessary. But do check with your local code requirements.

Haha. But seriously, I think they'd just smash down the door. When I do make this into a functional product, and if I ever do decide to sell it, I will 100% get it checked over by a lawyer and make sure it complies/exceeds all requirements and industry standards. The last thing that I (and probably the vast vast majority of designers) want is for my system to hurt or kill somebody.

[Whales]

The last thing that I (and probably the vast vast majority of designers) want is for my system to hurt or kill somebody.

Gah!  The last thing you want would be to have a lawyer check your design ;P

HalfSpace, it seems that most lipo batteries used today are 3.7v per cell. It would probably be best to just have a single cell and boost it up to the required voltage with a dc-dc converter.

Keep in mind how low you are going to discharge them to.  2.7 to 2.9V is a common bottom-end range.

I read what you wrote as saying that I should connect both the charging circuitry and the load directly to the battery and simultaneously charge and discharge the battery - however I may be wrong, but if that is what you meant then this is why I am hesitant: from what I've seen on the internet, when you both simultaneously charge and discharge a battery (in this case charging the battery from 48v PoE being stepped down to the charging voltage and being discharged by the Arduino and door striker) it does work; however it is seen as a bit of a hack as you are constantly running the battery, degrading its life and performance more rapidly over time.

One solution: a diode on each power source ( before the final voltage regulation) to isolate them from each other.  Something more efficcient could also probably be done with a pair of p-fets.

Regarding the Arduino's voltage regulator, the idea of wasting the 7v (12v input - 7v from volt reg to yield 5v for Arduino) sorta annoys me . However, I have been thinking recently () that it would be a good tactic to use a dc-dc converter to do the heavy lifting and bring whatever voltage (probably 12v) down to about 7v and then run it through a voltage regulator and get a much cleaner 5vdc.

Keep in mind MCs use only a handful of mA.  Depending on your choice/design of switchmode and the desired voltages in & out, you might find a linear DC-DC regulator is more efficient than a switchmode DC-DC one. 

Whales, I plan on using PoE in accordance with 802.3af, so it would be running at a nominal 48v. Thats an interesting approach. I was initially going to do that and design a little logic circuit or something and then use some MOSFETS to switch all of the power. I don't really think that it would be too difficult. The thing is that I began to read up on these PMIC's and got swayed a bit, they seem perfect for my situation but I just get lost in their data-sheets. I will make a circuit that should do the job but I would still love to see how to use one of these little power management ic's.

If you're going to market and mass-produce the product, you'll probably find one of the integrated solutions to be cheaper than a pile of discretes doing the same job.  None the less, do what you enjoy, and try both if you think you'll get more out of it that way

Thats a good idea to place a decent capacitors across the power rail (btw, would this be an example of a decoupling capacitor or is there another name for it?).

Thanks
Anti-browning cap? "Stops micro-controllers shitting themselves (TM)"